From mboxrd@z Thu Jan 1 00:00:00 1970 From: Thomas Petazzoni Date: Sun, 18 Sep 2011 19:14:44 +0200 Subject: [Buildroot] mdev problem in 2011.08 ? In-Reply-To: <4E76200B.40608@openwide.fr> References: <4E75EBBF.9080203@openwide.fr> <20110918162842.02150727@skate> <4E76200B.40608@openwide.fr> Message-ID: <20110918191444.1b721c91@skate> List-Id: MIME-Version: 1.0 Content-Type: text/plain; charset="us-ascii" Content-Transfer-Encoding: 7bit To: buildroot@busybox.net Le Sun, 18 Sep 2011 18:44:59 +0200, Pierre Ficheux a ?crit : > I test with QEMU + rootfs.cpio, does it matter? So you're using the rootfs.cpio as an initrd/initramfs ? If yes, then yes, it matters. The thing is that the kernel does not automount the devtmpfs filesystem when an initramfs is present. So, in Buildroot, we have an "init" wrapper script in fs/initramfs/init which mounts devtmpfs and does some other things before starting the real init process. However, this script is only used when you generate an initramfs bundled into a kernel image generated by Buildroot. If you generate a .cpio filesystem, this trick is not applied. However, Arnout Vandecappelle has recently contributed 2 patches that fix this by refactoring the initramfs and cpio support in Buildroot. See: From: "Arnout Vandecappelle (Essensium/Mind)" To: buildroot at busybox.net Subject: [Buildroot] [PATCH 1/2] cpio: fix boot with dynamic /dev Date: Tue, 6 Sep 2011 23:16:08 +0200 From: "Arnout Vandecappelle (Essensium/Mind)" To: buildroot at busybox.net Subject: [Buildroot] [PATCH 2/2] fs/initramfs: refactor with fs/cpio Date: Tue, 6 Sep 2011 23:16:09 +0200 The PATCH 1/2 should fix your problem, but on the long term, I think PATCH 2/2 is the way to go. If you have the time to test and Ack those patches, it would be nice! Regards, Thomas -- Thomas Petazzoni, Free Electrons Kernel, drivers, real-time and embedded Linux development, consulting, training and support. http://free-electrons.com